AMD’s New AI Chip Launch: Competing with Nvidia’s Blackwell

AMD has recently launched its new artificial intelligence (AI) chip called the ‘Instinct MI325X’, designed to compete directly with Nvidia’s data center graphics processors, or GPUs. These GPUs are essential for powering large-scale AI models and data centers. AMD is expected to begin production of this new chip within this year.

If AMD’s AI chips gain traction among developers and cloud giants as a viable alternative to Nvidia’s products, it could pressure Nvidia to adjust its pricing strategy. Last year, Nvidia’s GPUs saw a surge in demand, allowing the company to achieve an impressive 75% gross margin, securing its position as a market leader.

The Growing Need for Large Data Centers

Advanced generative AI models like OpenAI’s ChatGPT require powerful GPUs and large data centers to perform complex processing tasks. This demand has fueled the growth of the AI chip market, where Nvidia currently dominates. However, AMD is making significant strides to capture a larger share of this market. AMD’s CEO, Lisa Su, stated, “The demand for AI is growing faster than anticipated. Investments are scaling up everywhere.”

AMD’s Clients and MI325X Pricing

While AMD has not yet disclosed the names of new clients for its Instinct GPUs, it previously announced that major cloud providers like Meta and Microsoft already use its AI GPUs. Additionally, OpenAI employs some AMD chips for certain applications. The pricing details for the MI325X have not been released at this point.

Competing with Nvidia’s Blackwell Chips

AMD’s MI325X chip is set to go head-to-head with Nvidia’s upcoming Blackwell chips, expected to start shipping early next year. Investors interested in the booming AI sector may find AMD’s latest data center GPUs an attractive opportunity. So far in 2024, AMD’s stock has risen by only 20%, while Nvidia’s stock has surged over 175%.

Market experts estimate that Nvidia currently holds over 90% of the market share for AI chips in data centers. Recently, AMD’s stock dropped by 4%, while Nvidia’s saw a 1% increase.

Challenges AMD Faces

One of AMD’s biggest challenges in gaining market share is Nvidia’s proprietary programming language, CUDA, which has become the standard among AI developers. In response, AMD announced improvements to its ROCm software to make it easier for developers to transition their AI models to AMD chips.

Competing with Intel

AMD’s primary business has been central processors (CPUs). In July, the company reported that its data center revenue more than doubled in the June quarter, reaching $2.8 billion, with AI chips contributing $1 billion.

According to AMD, it holds around 34% of the market share in data center CPUs, which is still less than Intel’s dominance with its Xeon chip line. To strengthen its position, AMD has launched a new CPU line called EPYC 5th Gen, offering a variety of configurations, ranging from 8-core chips priced at $527 to powerful 192-core chips for supercomputers, priced at $14,813 each.